基于LSTM的沪深300ETF期权定价模型研究  

Research on Option Pricing Models of CSI 300ETF Based on LSTM

作　　者：赵可景 张金良[1] 朱怡梦 ZHAO Kejing;ZHANG Jinliang;ZHU Yimeng(School of Mathematics and Statistics,Henan University of Science and Technology,Luoyang 471000,Henan,China)

机构地区：[1]河南科技大学数学与统计学院,河南洛阳471000

出　　处：《山西师范大学学报（自然科学版）》2022年第3期31-38,共8页Journal of Shanxi Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(51675161).

摘　　要：将深度学习算法中的长短记忆神经网络(LSTM)引入期权定价的研究中.构建了沪深300ETF看涨期权和看跌期权的LSTM定价模型,进行了实证分析,并和BP神经网络模型的预测结果进行了对比.结果表明,LSTM神经网络模型的预测精度与深度学习的训练次数有关,且LSTM期权定价模型的预测效果要优于传统的BP模型.In this paper,the neural network of long and short time memory(LSTM)in deep learning algorithm is introduced into the research of option pricing.The LSTM pricing models of CSI 300ETF call option and put option are constructed,the empirical analysis is made,and the prediction results are compared with those of BP neural network model.The results show that the prediction accuracy of the LSTM neural network model is related to the training times of deep learning,and the prediction effect of the LSTM option pricing model is better than the traditional BP model.

关 键 词：期权定价 长短记忆神经网络(LSTM) BP神经网络 300ETF期权 

分 类 号：O211.6[理学—概率论与数理统计] F830.9[理学—数学]